A particle swarm optimization based multilateration algorithm for UWB sensor network

In this paper, a particle swarm optimization (PSO) based multilateration algorithm is presented for a UWB communications based sensor network. The particle swarm uses simple operators and is a bottom-up approach for identifying the location in a 2D space. Hence the PSO uses less energy. For comparison we present two alternative approaches traditionally used for this problem. The first one is (a) traditional iterative least square algorithm, (b) a one step simple least square solution. With respect to least squares, PSO results in slightly less error than the traditional iterative least square approach making. However, it is computationally inexpensive making it a good choice for a wireless network of small devices. The PSO multilateration algorithm really improves localization error over the one step least square algorithm. The new algorithm can replace the traditional algorithm in different applications.

I. Introduction

Advances in Ultra-Wideband (UWB) and ASIC technology have made it possible to build location-aware ad hoc sensor networks using inexpensive nodes consisting of a low power processor, a modest amount of memory, a wireless network transceiver and a sensor board. A wide range of applications such as personnel location and tracking for emergency responders, environmental monitoring, warehouse inventory tracking, vehicle locationing, building security, control of home appliances and military applications require accurate locationing systems [1]. GPS does not work at the locations of these underground and indoor applications. The designs of such location-aware networks typically require the capability of peer-to-peer ranging or a distance measurement. A UWB based locationing used in this paper is more beneficial under such circumstances. [2]. Relevant work has been carried out to investigate various ranging techniques for UWB, such as received signal strength intensity (RSSI), angle of arrival (AOA), time of arrival (TOA) and time difference of arrival (TDOA) [3]. Among these, TOA and TDOA are proven to be relatively accurate due to the excellent time resolution robustness in dense multipath environments of the UWB signals [4]. In line-of-sight (LOS) scenarios for UWB TOA-based ranging and positioning applications, accuracies can be in the order of tens of centimeters. The technique has been successfully employed in the UWB precision assets location system developed by Time Domain [5] [6].